search

MiraldiLab / maxATAC

Transcription Factor Binding Prediction from ATAC-seq and scATAC-seq with Deep Neural Networks
Apache License 2.0
25 stars 8 forks source link

--agg options for benchmark.md #86

Closed emiraldi closed 2 years ago

emiraldi commented 2 years ago

For the --agg argument, what are the options beyond the default "max"? Please update to include "sum" or "mean". (Note ranking by sum and the mean will be the same, so, if both sum and mean are arguments, you might want to update the codebase to remove the redundant option. I think having both sum and mean will confuse people!)

tacazares commented 2 years ago

The current options are max, coverage, mean, std, and min.

https://github.com/deeptools/pyBigWig#compute-summary-information-on-a-range

I will update the docs.

emiraldi commented 2 years ago

I don't understand how std would be used to rank or why that would make sense (omit unless I'm missing something). Is coverage % bp in region with a nonzero TFBS prediction or does it depend on a threshold? Please describe how these work, because it's not clear to me why we would want to allow some of these options. Maybe coverage makes sense for TF motif scanning but not maxATAC TFBS? We should only list the options that make sense with precision-recall (not all options that won't give an error with pyBigWig compute summary info).